Summary Basis of Decision for Jakavi

The Canadian regulatory decision on the clinical pharmacology, and clinical efficacy and safety studies was based on a critical assessment of the Canadian application/submission. Under the provisions of the Draft Guidance Document: The Use of Foreign Reviews by Health Canada, Method 3 was used whereby, during the review of the current submission, the foreign reviews completed by the United States Food and Drug Administration (FDA) and the European Union's centralized procedure European Medicines Agency (EMA) were used as an added reference in the assessment of Jakavi.

3.3.1 Pharmacodynamics

The STAT3 transcription factor is directly phosphorylated by JAKs in response to cytokine stimulation and therefore can be used as a pharmacodynamics (PD) marker for JAK inhibition. Following fasting, oral, multiple-dose administration, ruxolitinib demonstrated dose- and time-dependent inhibition of cytokine-induced STAT3 phosphorylated (p) with maximal inhibition occurring 1 to 2 hours after administration for all doses, coincident with the maximum plasma concentration (C_max). Maximal inhibition of STAT3p ranged from approximately 40% at the lowest dose (5 mg) to greater than 90% inhibition at the highest dose (200 mg). Levels of STAT3p returned to control levels by 24 hours in all treatments examined. The cytokine-induced STAT3p was inhibited by ruxolitinib with an IC₅₀ of 224 nM.

In a double-blind, placebo-controlled, crossover electrocardiogram study in 49 healthy volunteers, there was no indication of a QTc prolonging effect of ruxolitinib at single doses of 25 mg and 200 mg.

3.3.2 Pharmacokinetics

Absorption

Ruxolitinib is a Class 1 molecule under the Biopharmaceutical Classification System, with high permeability, high solubility and rapid dissolution characteristics. The drug was absorbed rapidly with a peak plasma concentration approximately 1 hour after oral administration. The plasma concentrations declined in a monophasic or biphasic fashion. The absorption of ruxolitinib was 95% or greater. The mean C_max and total exposure (AUC) were dose-proportional and linear pharmacokinetics (PK) were demonstrated over a single dose range of 5 mg to 200 mg. It was demonstrated that ruxolitinib does not accumulate and that it has time-independent PK.

The food effect was evaluated in healthy volunteers. The PK results demonstrated that there should be no clinically significant differences when ruxolitinib is administered under fed or fasted states in patients.

Distribution

Ruxolitinib was highly bound to human serum albumin with an unbound fraction of 3.3%. The volume of distribution at steady-state was 53-65 L in myelofibrosis patients.

Metabolism

CYP3A4 is the major enzyme responsible for the metabolism of ruxolitinib. Ruxolitinib is metabolized mostly by oxidation and the two major metabolites were identified as M18 and M16. The metabolites do retain pharmacological activity. The parent compound was the predominant entity in circulation representing approximately 60% of the drug-related material in circulation.

Excretion

The recovered parent compound was 74% from urine and 22% from faeces, thus urine is the main route of excretion. Less than 1% was unchanged drug, in concordance to the observation that ruxolitinib is eliminated almost completely by oxidative metabolism with a terminal half-life of approximately 3 hours.

Drug-Drug Interaction Studies

The PK parameters (C_max, AUC, and clearance) were strongly influenced when ruxolitinib was administered concomitantly with ketoconazole, a strong CYP3A4 inhibitor, therefore dose adjustments were recommended. There were no significant PK changes observed when ruxolitinib was co-administered with erythromycin, a moderate CYP3A4 inhibitor, therefore dose adjustments were not recommended with the co-administration of moderate to mild CYP3A4 inhibitors.

The PK effects of the co-administration of ruxolitinib with rifampicin, a strong CYP3A4 inducer, were minor and the PD parameters did not change. Therefore, no dose alterations were recommended when strong CYP3A4 inducers are co-administered.

Special Populations

A hepatic impairment study showed that a single 25 mg dose of ruxolitinib administered in patients with mild, moderate, and severe hepatic impairment, increased the AUC 1.87, 1.28, and 1.65 times, respectively. The terminal elimination half-life (t_½) was 1.6, 1.5, and 1.8-fold prolonged for the mild, moderate, and severe hepatically impaired patients, respectively. The clearance (CL/F) was approximately 2-fold decreased for the three groups. It was recommended that patients with mild, moderate, or severe hepatic impairment receive a lower dose of ruxolitinib based on their platelet count.

A renal impairment study showed no major differences in the PK or PD of ruxolitinib when administered in patients with varying degree of renal impairment (mild, moderate, or severe). However, the 8 active metabolites contributed to PK parameter changes evaluated in patients with renal impairment. These PK changes translated into an increased PD combined effect (parent + metabolites) of 117%, 123%, 134%, 153%, 212% and 192% for normal, mild, moderate, severe, end-stage renal disease (ESRD) with dosing after dialysis, and ESRD with dosing before dialysis, respectively. Based on the observed changes of the PK and PD due to ruxolitinib's active metabolites, the final recommendations propose dose reductions in patients with moderate and severe renal impairment. Even though there was a limited data of ESRD patients on dialysis, the results showed significant PK and PD changes, thus the final recommendations were to alter the administration schedule of ruxolitinib.

3.3.3 Clinical Efficacy

The efficacy and safety findings were derived from two pivotal Phase III studies and one supporting Phase II study. COMFORT-I [number of patients (n) = 309] was a randomized, double-blind, placebo-controlled Phase III study in patients with Primary Myelofibrosis (MF), Post-Polycythaemia Vera Myelofibrosis (PPV-MF) or Post-Essential Thrombocythaemia-Myelofibrosis (PET-MF). COMFORT-II (n = 219) was a randomized, open-label, efficacy and safety Phase III study of JAKAVI tablets compared to best available therapy (BAT) in patients with PMF, PPV-MF or PET-MF.

The starting dose of Jakavi was based on platelet count. Patients with a platelet count between 100,000 and 200,000/mm³ were started on Jakavi 15 mg twice daily (BID) and patients with a platelet count >200,000/mm³ were started on Jakavi 20 mg BID. A standardized dosing paradigm was used to determine dose adjustments for safety and efficacy so that each patient was titrated to their most appropriate dose. Doses were not to exceed 25 mg BID. The recommended dose titration has been fully justified by the results of the pivotal studies and by one supporting Phase II study.

The comparator groups were relevant. In COMFORT-I, a placebo was used as the comparator while in COMFORT-II, BAT was used as the comparator. Although, none of the drugs used in the BAT group were approved for use in the treatment of MF in Canada, the results of COMFORT-II are supportive of the other pivotal study which compared Jakavi to placebo.

Primary Endpoints

The primary efficacy endpoint in the pivotal studies was the proportion of patients achieving ≥35% reduction in spleen volume from baseline to Week 24 (COMFORT-I) or to Week 48 (COMFORT-II) as measured by Magnetic Resonance Imaging (MRI) or Computerized Axial Tomography (CAT).

Both pivotal studies met their primary endpoint. A statistically significant higher proportion of patients achieved ≥35% reduction in spleen volume with Jakavi compared to controls at Week 24 [COMFORT-I; 41.9% versus (vs.) 0.7%] or at Week 48 (COMFORT-II; 28.5% vs. 0%).

The response rates were lower in male patients compared to female patients, suggesting that the efficacy of ruxolitinib could be lower in male patients. However, a higher starting dose in male patients is not recommended because it may increase the risk of developing Jakavi-associated toxicities. Any potential difference in efficacy in male patients can be managed by the titrating approach while preserving these patients from excessive toxicity.

The response rates were smaller in patients with a starting dose of 15 mg compared to patients with a starting dose of 20 mg. However, the administration of a 20 mg starting dose in patients with a platelet count of 100,000/mm³ to 200,000/mm³ is not recommended since it may result in an unacceptable risk-benefit profile in these patients.

Secondary Endpoints

Secondary endpoints in COMFORT-I included the duration of maintenance of a ≥35% reduction from baseline in spleen volume, the proportion of patients who had ≥50% reduction in total symptom score from baseline to Week 24 as measured by the modified Myelofibrosis Symptom Assessment Form (MFSAF) v2.0 diary, the change in total symptom score from baseline to Week 24 as measured by the modified MFSAF v2.0 diary, and overall survival (OS).

In COMFORT-II, the secondary endpoints were the proportion of patients achieving a ≥35% reduction of spleen volume measured by MRI or CAT from baseline to Week 24 and the duration of maintenance of a ≥35% reduction from baseline in responding patients.

In COMFORT-I, a significantly larger proportion of patients in the Jakavi group achieved a ≥50% improvement from baseline in the Week 24 total symptom score compared with the placebo group [45.9% and 5.3%, respectively, probability (p) <0.0001 using the Chi-Squared test], as measured by the modified MFSAF v2.0 diary.

COMFORT-II also met its key secondary endpoint. A statistically significant higher proportion of patients treated with Jakavi achieved ≥35% reduction in spleen volume comparing baseline to Week 24 (32% vs. 0% in Jakavi group and BAT group respectively).

No statistical difference in OS between the groups was observed in either pivotal study. An additional OS analysis was performed in COMFORT-I, which included the data from an additional follow-up period. Although the results from the follow-up analysis showed an OS improvement in the Jakavi group, there was no adjustment for repeated testing and it was not pre-specified in the protocol. These results are considered as exploratory.

In COMFORT-II, the median duration of maintenance of spleen volume reduction in Jakavi-treated patients was 48 weeks. The median duration of response in COMFORT-I was not reached as the majority of patients were still responding at the data cutoff.

The median time to ≥35% reduction in spleen volume was 12.29 weeks in COMFORT-II, which may have been overestimated due to the lack of spleen assessment before 12 weeks.

The clinical studies contain no evidence that Jakavi can modify the course of myelofibrosis. The goal of therapy with Jakavi appears to be the control of some of the MF symptoms, namely splenomegaly.

There is no evidence demonstrating the clinical benefit of the Jakavi treatment in asymptomatic myelofibrosis patients (i.e., without splenomegaly or constitutional symptoms).

In conclusion, the data from the pivotal studies demonstrated that Jakavi can significantly reduce splenomegaly in patients with PMF, PPV-MF or PET-MF, including patients with massively enlarged spleens. Jakavi was able to halt the progression of splenomegaly in almost all patients. Patients also had significant clinical benefit in the reduction of important symptoms of MF as assessed by the modified MFSAF v2.0 diary. However, it is unclear whether the reduction of the symptoms of MF is due to a reduction of the splenomegaly or to a direct effect of Jakavi on those symptoms. Consequently, it is currently unclear whether MF patients (symptomatic or asymptomatic) without splenomegaly may benefit from the Jakavi therapy.

The proposed clinical population in the indication was too broad since it included all patients with MF regardless of the presence of splenomegaly, and the efficacy results did not support the sponsor's proposed indication. The targeted population should be defined by the symptoms to be treated, in occurrence splenomegaly and its associated symptoms. Therefore, the indication should specify that Jakavi is for the treatment of splenomegaly and/or its associated symptoms in patients with PMF, PPV-MF or PET-MF.

3.3.4 Clinical Safety

The safety profile of Jakavi in patients with myelofibrosis was derived from 589 patients treated in two pivotal Phase III studies and one Phase II supporting study.

In the two pivotal studies COMFORT-I and COMFORT-II, 301 patients had a median duration of exposure to Jakavi of 10.8 months (range 2 weeks to 19.5 months). The majority of patients (68.4%) were treated for at least 9 months. Of the 301 patients, 111 (36.9%) had a baseline platelet count between 100,000/mm³ and 200,000/mm³, and 190 (63.1%) had a baseline platelet count >200,000/mm³.

Study results suggested that individual dose titration within the 10 to 25 mg range offered the best balance of efficacy and safety for an individual. The dose that was ultimately tolerated appeared to be mainly influenced by the baseline platelet count. There is no experience with the administration of Jakavi to patients with platelet counts <100,000/mm³ at the start of treatment. Grade 3 and 4 thrombocytopenia can usually be avoided by dose interruption or reduction.

In the Jakavi group, the most common adverse drug reactions (ADRs) were related to myelosuppression, namely thrombocytopaenia and anaemia. This may be due to the disruption of the JAK/STAT pathway by Jakavi, which is an important signalling pathway in haematopoiesis. Anemia and thrombocytopaenia were dose-related effects.

The most frequently reported AE leading to dose interruption or discontinuation of study medication was thrombocytopaenia. The most frequent AE leading to dosage reduction was thrombocytopaenia/decreased platelet count. In both pivotal studies, dose reductions due to anaemia occurred in 5.3% of Jakavi-treated patients.

Other ADRs reported with a higher frequency in the Jakavi group compared to control groups (in both pivotal studies) were: neutropaenia; ecchymosis; pyrexia; headache; weight increased; haematoma; dizziness; urinary tract infection; hypercholesterolemia; herpes zoster; palpitations; bradycardia/sinus bradycardia; angina pectoris/angina unstable; and flatulence.

Haematoma and haemorrhage are adverse events of interest due to the mechanism of action of the drug. Although the proportion of patients with at least one bleeding event was higher in the subgroup of patients who received Jakavi and acetylsalicylic acid compared to patients who received acetylsalicylic acid in the control groups, there is insufficient evidence to suggest that concomitant use of aspirin increases the risk of bleeding events in Jakavi patients. In fact, data shows that there was a lower risk of bleeding events in patients who received Jakavi and acetylsalicylic acid compared to patients who received Jakavi and no anticoagulant (33.7% vs. 51.0%). Bleeding events accounted for nine deaths (1.1%) in the Jakavi group (subdural haematoma, retroperitoneal haemorrhage, 2 gastrointestinal haemorrhage, 2 cerebral haemorrhage, 3 other haemorrhage) and 2 deaths (0.7%) in the placebo group (gastrointestinal haemorrhage and subdural haematoma). Therefore, there is insufficient evidence to support a boxed warning or the addition into the Adverse Reactions section of information related to an association between fatal haemorrhage and the concomitant use of Jakavi and acetylsalicylic acid.

The proportion of patients who had a serious adverse event (SAE) was higher among patients in the placebo group (35.1%) compared with the Jakavi group (27.7%) in COMFORT-I. In COMFORT-II, SAEs were reported for a similar proportion of patients in the two groups: 30.1% and 28.8% of patients in the Jakavi group; and the BAT group, respectively.

Increased levels of alanine transaminase (ALT) and aspartate aminotransferase (AST) were more frequent in the Jakavi group compared to the control groups. Most cases of hepatotoxicity in the Jakavi group were of mild severity and appeared to be manageable. Severe drug-induced liver injury does not seem to be associated with Jakavi.

Treatment with Jakavi was associated with statistically significant PR interval prolongation and decrease heart rate. Cardiac disorders were observed at a higher frequency in the Jakavi group compared to the control groups (COMFORT-I; 18.9% vs. 13.9% while in COMFORT-II; 19.9% vs. 11%). Bradycardia/sinus bradycardia, palpitations, and angina pectoris/angina unstable, occurred at a higher rate during treatment with Jakavi in both pivotal studies. The association of cardiac failure/cardiac failure congestive, hypertension/hypertensive crisis, and myocardial infarction/acute myocardial infarction with Jakavi is less clear because they were observed at a higher rate in only one of the pivotal studies.

In contrast to COMFORT-I and the QT interval study in healthy subjects, a statistically significant increase from baseline in the QTcF interval (QT corrected using Fridericia's formula) was observed in COMFORT-II at Week 4 and Week 24, but not at Week 12 and Week 48. QTcF interval prolongation in the 450-480 ms range was observed in 7 patients (4.8%) in the Jakavi group and 0 patients in the BAT group, prolongation between 480-500 ms was seen in 1 Jakavi-treated patient (0.7%) and 2 BAT-treated patients (2.7%). However, the actual effect of Jakavi on QTc interval is currently unclear because the submitted studies were not designed to thoroughly assess the effect of Jakavi on QTc interval.

There is no experience with the administration of Jakavi to patients with absolute neutrophil count (ANC) ≤1,000/µL. Resolution of decreased levels of neutrophils occurred without dose adjustment in the majority of patients. Few patients had neutrophil counts that required dose interruption per the protocol (ANC <500/μL).

Overall, infection rates and infections associated with immunosuppressive drugs in particular were not more frequent in the Jakavi groups compared to the control groups.

In COMFORT-I, more patients moved from transfusion independence to dependence in the Jakavi group, suggesting that Jakavi increases the likelihood of transfusion requirement. This finding could not be confirmed in COMFORT-II due to the large number of patients with missing transfusion status.

Following interruption or discontinuation of Jakavi, symptoms of myelofibrosis may return over a period of approximately 1 week. It has not been established whether abrupt discontinuation of Jakavi contributed to these events. This is appropriately captured in the Product Monograph.

Animal studies have shown that ruxolitinib is embryotoxic and foetotoxic (in rats and rabbits). The potential risk of teratogenicity for humans is unknown. The exposure margins in the non-clinical studies compared to the highest clinical dose were low and the results were therefore of limited relevance for humans. In addition, the presence of ruxolitinib in human semen cannot be ruled out. The level of ruxolitinib in semen that would present no increased risk to the embryos is currently unknown. Appropriate warnings and precautionary measures are in place in the Product Monograph to limit the exposure of the embryos to ruxolitinib.